I have a table with more than 50,000 records in SQL server 2000 without any index or even primary key on any column. Now i have identified two columns which are together unique ( say- empid, projectid) for the table. Now if i create a clustered index on the above mentioned columns, then my table scans are getting replaced by clustered index seeks in the execution plan for queries but the situation is that frequntly new records might need to be inserted inbetween the sorted clustered records (as when new projects are associated to an existing empid). So it might result into page splits. So how advisable is it to have a clustered index in these kind of situations where the clustered indexed column is not monotonically increasing. Does creating one with a lower fill factor and pad index should be considered. What should be the fill factor and pad index values. What about creating a clustered index on a column that is not unique and with similar situation as above. As there will be no nonclustered indexes so can this be considered. Thanks in advance.